XgCPred: Cell type classification using XGBoost-CNN integration and exploiting gene expression imaging in single-cell...

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) enables cellular and developmental biology research.
• Accurate cell type classification is crucial for understanding tissue composition and disease origins.
• Current methods face challenges with data variability, aggregation, and high dimensionality.

Purpose of the Study:

• To develop a novel computational approach for accurate cell type classification in scRNA-seq data.
• To address limitations in existing methods for handling complex and large-scale scRNA-seq datasets.
• To improve the reliability of cell annotations for downstream biological research.

Main Methods:

• XgCPred combines XGBoost and Convolutional Neural Networks (CNNs).
• It uses an imaging representation of gene expression based on KEGG BRITE hierarchy.
• This approach leverages CNNs for spatial hierarchy detection and XGBoost for large-volume data processing.

Main Results:

• XgCPred demonstrated superior performance across diverse scRNA-seq datasets.
• The method achieved high accuracy and precision in cell type annotation, with near-perfect scores in some cases.
• Results highlight XgCPred's ability to manage data variability and heterogeneity effectively.

Conclusions:

• XgCPred provides dependable and accurate cell type classification for scRNA-seq data.
• The approach offers a scalable and potent solution for growing dataset sizes and complexity.
• XgCPred advances genomic research, aiding in biological discovery and disease detection by improving computational efficiency and generalizability.